|
2004-09-03 21:29
|
fs2_open_3_6 2004-09-04 01-27-33-20.JPG (82,036 bytes) |
|
|
Have you gotten anywhere with this? Does 3.6.5 beta work any better? |
|
|
Still green with the 3.6.5 beta, now with Cats 4.12. |
|
|
Does it for me aswell, Omega Cats 4.12 beta |
|
|
Installed 3.6.5 final today, fired up FS2 and noticed that everything was 16bit eventhough I had 1024x768x32 selected in the launcher. So, I exited and switched to 1024x768x16, launched again and everythung was 32bits and the greenery is gone. |
|
|
Goober's 20050125 build has fixed the problems, using cats 5.1, btw. |
|
|
Hmm. I've made a few changes to the OpenGL code but nothing that directly relates to the green screen problem. Can others please verify that this is fixed. I'm still leaning towards a driver issue since I've never seen this myself and have tested it on nearly 20 different systems. |
|
|
I can verify that this problem is present in your 20050212, the 20050103 build, and 3.6.5. (using 5.02CAT's - i also had this problem on 4.12 cats as well) On top of the Green Screen, the initial loading screen is always garbled, or shows the last image seen before exiting FS2. edited on: 02-27-05 00:05 |
|
|
I've seen this as well, though the spash screen seems to be getting saved somewhere that gets called up whenever OGL gets initialized. The green has gone away though in recent builds (using Catalyst 5.2). If it's still a problem under Cat 4.x, I'd think that upgrading is in order. |
|
|
I get this too, but it isn't necessarily green. For me it is just random garbage. Updating drivers doesn't help. ATI Radeon IGP 340M (an integrated laptop chip). |
|
|
Is there any difference in how it behaves when you run in 16-bit vs. 32-bit color modes? Or in a window rather than fullscreen? edited on: 03-02-05 21:54 |
|
|
Just tested this using build T-20050103 & 5.02 cats on 9800 32-bit fullscreen : Green screen / Loading screen garbled 32-bit windowed : No problems 16-bit fullscreen : No problems 16-bit windowed : No problems |
|
|
Ooh, now we're getting somewhere. I'm going to try a few different things and see what I come up with. Looks like something with the pixel format, I just need to figure out exactly what and get around it without breaking anything else. I do see a couple of possible issues though so I'll try and post a build here for everyone to test as soon as I can. EDIT: I don't hold out much hope but give this a try: http://icculus.org/~taylor/fso/willrobinson/20050302-win32r.zip edited on: 03-03-05 00:15 |
|
|
Never mind, thought this was new without comments, assigning to taylor. edited on: 03-13-05 14:25 |
|
|
OK, I just tested this out and here's what happened. -32-bit fullscreen : Loading screen fine BUT green screen shows up when you press escape to exit and you get the exit confirmation window. -32-bit windowed : No problems -16-bit fullscreen : Same as 32 bit fuill screen. Loading screen fine BUT green screen shows up when you press escape to exit and you get the exit confirmation window. -16-bit windowed : No problems |
|
|
Alright I'll try a few other things and report back. I'm still trying to find a way to consistently reproduce the problem. Have done it a couple of times but not enough to get a fix. |
|
|
I just spotted this under D3D, as well as OGL. fs2_open_r_20050414.exe -spec -glow -jpgtga -d3dmipmap -nomotiondebris -2d_poof -snd_preload -env -alpha_env -fps -fov 0.39 (Goobs latest) Radeon X800 Pro AMD64 3000+ 1GB RAM 17 inch TFT. Windows XP Home Also, in my case it looks more blueish. RGB: 0,237,242 according to PSP7 after a screenie. Spotted the same thing on an older build on another computer, same colour. Pentium 4 1,2Ghz 384MB Rambus RAM Some kind of geforce. 15 inch TFT. windows ME. I can't really test on either comp, because on is my brothers and the other one is my dad's. |
|
|
If it just showed up in D3D can you go back and tell me what the last build was that worked? If a recent code change caused this for D3D then maybe that will help track down the OGL cause too. |
|
|
I'm not seeing this any more at all. The most that happens in OGL is that the background behind a popup is blown up if I'm playing at a nonstandard res (and D3D is just screwed up with higher res) but the background is always visible in some form. |
|
|
Can't test on either machine since neither is mine. I only played with the new stuff for a few minutes. |
|
|
@StratComm: I thought that I fixed the OGL blow up problem already. Would only be in builds that came out after 03/05/05 though. If you are using a newer build and still have the problem let me know. |
|
|
You're right, it's gone in the newest builds. I've not been up to date with builds for about a month prior to the current wave, and I hadn't noticed one way or the other yet. But everything released this month looks good. |
|
|
Update: Back as of Goob's builds of 4-23 and 4-25. |
|
|
Howabout more recent builds? |
|
|
Got a possible fix for this but haven't tested it on a computer with the problem yet. Check my next build (few days away) and see if it's fixed now. Still not really sure why this fix would work, but after a lot of Googling, it could get the job done. |
|
|
I made several changes to my latest build to try and fix this so give it a try and let me know if it's better or not. http://dynamic.gamespy.com/~freespace/forums/showthread.php?s=&threadid=33250 |
|
|
Two people say it's ok now so it's either a fluke or actually fixed. If anyone still sees this the just reopen the bug. And just so we can keep track of what changed: GL_TEXTURE_2D is disabled before the glReadPixels call and re-enabled right after. After a lot of google searching and reading on opengl.org it appeared that some people were getting certain solid colors with the gl*Pixel operations when GL_TEXTURE_2D was enabled. I'm still not 100% sure this is all it takes to fix this but since I've only heard positives and no negatives I'll assume that it is. Fixered. |
|
|
The greenery is back with Goober's last two builds, 20050728 and the one before it, both with Cat 5.7 and 5.8. |
|
|
Does it happen on Sticks's build? |
|
|
yes |
|
|
I'm opening this bug again, and keeping it assigned to Taylor since he was working on it. Sorry for adding more work to your list, but you seem to know the code. |
|
|
Ok I think this might be really fixed now (if warpcore allows me to commit any time soon). Screenshots appeared to messing up like this same thing so since that code is a bit easier to change I made a few tests there. It would seem like the ATI drivers don't like glReadPixels with a BGR format, but I have no concrete basis for that. It worked as far as screenshots go though. So, I changed the popup code to also use RGB (instead of BGRA), swap by hand back to BGR, and I think it's happy now. And it uses less memory. Assuming that the server starts working and I can finally commit this change we'll see if it works. It will proably be in the very next CVS build. Let me know whether it works or not. |
|
|
I'm sorry but I can't test your fix, I'm currently stuck with a Radeon 8500 and it's not going to change anytime soon. |
|
|
??? The fix is for Radeon cards (all of them)... |
|
|
With WMCs 29/12 build i'm getting green background in pause and options (esc) screen. With ATI Radeon 9800Pro. |
|
|
Do screenshots work? |
|
|
Just tested it. Odd, with 22/12 CVS build i could take good screenshots and they were saved in the mod folder but now they seem to be saved in totally different folder (freespace2/screenshots) and are all green. With WMCs fs2_open_r-20051229.exe (standard build). |
|
|
Ok, the code that previously fixed screenshots is back in CVS now. Same basic thing done for popups, though I'm sure it going to look strange unless you run in a window rather than fullscreen. If it works (single-color problem gone) in the next available build then please run a debug build, go to the mission simulator, exit, and attach the debug log here. I have an idea of what the problem might be (I hope so anyway) and the debug log will have a line or two that I need to find out for sure. |
|
2006-01-07 09:48
|
|
|
|
I didn't see any green background in mission load screen (during the pop-up), in game with pause or in screenshots. fs.log has been added as an attachment. |
|
|
So it's working fine now then? The popups show the correct background? Please also double check it running in a window too since it will behave a little differently. |
|
|
No green backgrounds in popups, pause screens, or in screenshots. Tests were made with OGL 1024 x 768 x 32 (windowed & normal) and 640 x 480 x 32 (windowed only). So atleast for me it is looking pretty good. Background pics also seem to be the correct ones (pause screen shows a 'screenshot' from the moment game was 'paused'). |
|
|
Nice. Ok, I'll leave this bug open a little longer for any others to comment. If it looks fixed across the board (ie, it still works on NVIDIA cards) then I'll resolve it. I've got a mental note about the problem/fix now so if this crops up again then it shouldn't be too difficult to deal with. |
|
|
No bad news. Guess it's really fixed this time. :D Fixered. |
|
|
IT'S BACK!!!! |
|
|
Please someone good should look at this bug, I need to make screenshots for debugging 3.6.8.zeta |
|
|
I use catalyst 5.8 and when i run fs2_open_C05212005_d-P4 in fullscreen I experience the same bug, but when i run in a window and make screenshot it looks weird look at screen0034.jpg. I used opengl/cannot launch d3d/. MediaVPs are 3.6.8.zeta, on delta it was ok. Heres my cmd line: D:\Hry\Freespace2\fs2_open_C05212005_d-P4.exe -spec -glow -jpgtga -mipmap -nomotiondebris -2d_poof -missile_lighting -dualscanlines -orbradar -warp_flash -snd_preload -env -alpha_env -fps -window -fov 0.50 -ambient_factor 75 -no_warn I know it happens also on FS2_Open_KARA_20060517 maybe other builds too. I posted also my fs.log/shotman_fs.log/ Maybe it helps, good luck. edited on: 05-24-06 10:35 |
|
2006-05-24 14:31
|
screen0034.jpeg (35,023 bytes) |
|
2006-05-24 14:31
|
|
|
|
Shotman, try not to double-post. :p Duplicate posts deleted. |
|
|
Just a quick note that this is verified as once again fixed in private test builds. The fix won't hit CVS until some time this weekend though when I have a chance to commit my other code changes. |
|
|
Should be re-fixed in CVS. If anyone still problems with the next build then just reopen this and I'll take another look. Fixered. |
- Anonymous
